Scam Detector fix false positive (#1216)

* Added unit tests for false positives

* added feature so sus keywords can require startsWith instead of contains

^foo instead of just foo

* CR improvement

Author: Zabuzard

application/config.json.template

@@ -45,7 +45,7 @@
             "steamcommunity",
             "freenitro",
             "usd",
-            "earn",
+            "^earn",
             ".exe"
         ],
         "hostWhitelist": [

application/src/main/java/org/togetherjava/tjbot/features/moderation/scam/ScamDetector.java

@@ -114,7 +114,13 @@ private boolean containsSuspiciousKeyword(String token) {
         return config.getSuspiciousKeywords()
             .stream()
             .map(keyword -> keyword.toLowerCase(Locale.US))
-            .anyMatch(preparedToken::contains);
+            .anyMatch(keyword -> {
+                // Simple regex-inspired syntax "^foo"
+                if (startsWith(keyword, '^')) {
+                    return preparedToken.startsWith(keyword.substring(1));
+                }
+                return preparedToken.contains(keyword);
+            });
     }
 
     private boolean isHostSimilarToKeyword(String host, String keyword) {
@@ -140,6 +146,10 @@ private boolean isHostSimilarToKeyword(String host, String keyword) {
         return false;
     }
 
+    private static boolean startsWith(CharSequence text, char prefixToTest) {
+        return !text.isEmpty() && text.charAt(0) == prefixToTest;
+    }
+
     private static class AnalyseResults {
         private boolean pingsEveryone;
         private boolean containsSuspiciousKeyword;

application/src/test/java/org/togetherjava/tjbot/features/moderation/scam/ScamDetectorTest.java

@@ -26,10 +26,10 @@ void setUp() {
         ScamBlockerConfig scamConfig = mock(ScamBlockerConfig.class);
         when(config.getScamBlocker()).thenReturn(scamConfig);
 
-        when(scamConfig.getSuspiciousKeywords())
-            .thenReturn(Set.of("nitro", "boob", "sexy", "sexi", "esex", "steam", "gift", "onlyfans",
-                    "bitcoin", "btc", "promo", "trader", "trading", "whatsapp", "crypto", "claim",
-                    "teen", "adobe", "hack", "steamcommunity", "freenitro", "usd", "earn", ".exe"));
+        when(scamConfig.getSuspiciousKeywords()).thenReturn(Set.of("nitro", "boob", "sexy", "sexi",
+                "esex", "steam", "gift", "onlyfans", "bitcoin", "btc", "promo", "trader", "trading",
+                "whatsapp", "crypto", "claim", "teen", "adobe", "hack", "steamcommunity",
+                "freenitro", "usd", "^earn", ".exe"));
         when(scamConfig.getHostWhitelist()).thenReturn(Set.of("discord.com", "discord.media",
                 "discordapp.com", "discordapp.net", "discordstatus.com"));
         when(scamConfig.getHostBlacklist()).thenReturn(Set.of("bit.ly", "discord.gg", "teletype.in",
@@ -54,6 +54,18 @@ void detectsRealScam(String scamMessage) {
         assertTrue(isScamResult);
     }
 
+    @ParameterizedTest
+    @MethodSource("provideRealFalsePositiveMessages")
+    @DisplayName("Can ignore real false positive messages")
+    void ignoresFalsePositives(String falsePositiveMessage) {
+        // GIVEN a real false positive message
+        // WHEN analyzing it
+        boolean isScamResult = scamDetector.isScam(falsePositiveMessage);
+
+        // THEN does not flag it as scam
+        assertFalse(isScamResult);
+    }
+
     @Test
     @DisplayName("Can detect messages that contain blacklisted websites as scam")
     void detectsBlacklistedWebsite() {
@@ -227,4 +239,10 @@ private static List<String> provideRealScamMessages() {
                 "Urgently looking for mods & collab managers https://discord.gg/cryptohireo",
                 "Check this - https://transfer.sh/get/ajmkh3l7tzop/Setup.exe");
     }
+
+    private static List<String> provideRealFalsePositiveMessages() {
+        return List
+            .of("""
+                    https://learn.microsoft.com/en-us/dotnet/csharp/fundamentals/types/anonymous-types""");
+    }
 }